best way to handle switching in audio pre-amp

What is the best way to handle analog switching for line-level high-fidelity audio? I don't think I want to use mechanical switching because nice switches seem kind of pricey, and I might want to use a remote control in a future revision.

So, I guess I am looking for a good analog mux or switch which won't introduce distortion. For example, I looked at the SNL74LVC2G53, but it seems as though it would add too much distortion for a hi-fi application.

Can I just use an N-channel MOSFET with the gate pulled to +/- 15 Volts for on and off, respectively? Any particular part recommendations?

I think I remember Win recommending using two MOSFET's in series, but I think the idea was to float the gates at vs + 15V somehow. I don't want to do anything that tricky if I can avoid it. ;-)

Oh, I am planning on having +/- 15 volt rails for the analog circuitry, and I guess I'll need some 5 Volts for the digital circuitry. I'd like to keep the 5 Volts to, say, under 10mA, if possible.

--Mac

You can get integrated analogue switches. Siliconix's DG211 and 308 used to be classic workhorses but nothing beats using jfets.

Check Siliconix's J111 and J174 families. As widely used in pro-audio.

You'll have to devise your own driver circuitry though !

Graham

I would have recommended bistable relays, which do not consume power unless switching.

If you really want +/-15V supply and high level audio this part won't do. Not because of distortion but of breakdown voltage. BTW how high would be the distortion induced by an analog switch?

There is an intrinsic diode in most mosfets which would conduct even if the transistor is switched off.

see above to avoid getting the diode limit output voltage. And with the high levels you want, you will have to use 20V to be above Vp.

Why the high voltage if an amplifier needs less than 1Veff. for full power? maybe it would be better to buy a ready made preamp with your design skills?

--
ciao Ban
Bordighera, Italy

Thanks, Ban. I kind of suspected something like that must exist because my home thermostat uses relays (I can hear them click when the heat goes on or off) but it runs only from a small alkaline battery, and has been going strong for more than a year.

Oh, I'm sorry, my idea was to use +/- 15 Volt rails because the op-amp I want to use (OPA2134, which I discovered in a previous thread) works best with the high rails. I don't intend to have high level signals present on the output. I would have only line level signals, as I said in the original post.

Ah, yes. Good point. I even knew this, but somehow the significance of it slipped my mind. Thanks Ban.

My design skills are not as weak as you think. I don't do a lot of analog design, especially analog design with discrete transistors, but I have designed a few op-amp circuits and lots of digital stuff.

Thanks again, Ban.

--Mac

After I posted, I realized that I hadn't even considered relays. Thanks for recommending specific ones. I will have a look.

--Mac

I still use relays,

for example series 100 Ideal for data acquisition - Thermal EMF of 1pV or less!

or series 101, 74 logic compatable

martin

Serious error. All shortcuts have disappeared. Screen. Mind. Both are blank.

Thanks. I'll take a look. I especially appreciate that you listed specific parts.

--Mac

[snip]

[snip]

For the part mentioned above, the datasheet says 0.01% into 10k || 50 pF. This is at 10 kHz. Interestingly, they list the same distortion value for

1 kHz.

Now that I read it over and actually look at the test setup, I see that they are swinging to within 0.5 Volts of the rail. If I power it at

+/- 2.5 Volts and keep the load much higher than 10 k, I'm sure I will get better performance at realistic levels.

So maybe I can use a cheap mux afterall.

It uses so little power that I don't mind having to generate the

+/-2.5 Volt rails. And yes, I know I have to watch out that I don't drive the inputs beyond those rails.

By the way, if you can recommend a good op-amp for line-level buffering that will work well at +/- 5V, I would be very interested. In fact, I would probably just change my design over to using +/- 5V instead of +/-

1. I guess I would see 4 or 5$ per op-amp (at low volume) as the upper limit on price.

Anyway, thanks again, Ban! You seem to almost always have good advice for people interested in audio stuff.

--Mac

Mac wrote: ]

If you want to buffer (no gain) I would recommend the SSM2135 opamp, which is fully specd at +5V and can even drive 32ohm loads(headphones) with low distortion. For gain stages I recommend the LT1124. The op you mentioned before OPA2134 will also work on lower voltages. Just avoid the CMOS stuff, which tends to be noisy in the bass. To use cheap and small electrolytic capacitors (bridged with a small film cap) instead of huge film caps it might be suitable to have only a single supply, i.e. +12V. When switched on this gives a big bump tho, but an output relay can disconnect or ground the amp if power is not yet ok.

THX, have a look at my loudspeakers

--
ciao Ban
Bordighera, Italy